Human-artificial intelligence symbiotic reporting for theranostic cancer care

Abstract

Reporting of diagnostic nuclear images in clinical cancer management is generally qualitative. Theranostic treatment with 177 Lu radioligands for prostate cancer and neuroendocrine tumors is routinely given as the same arbitrary fixed administered activity to every patient. Nuclear oncology, as currently practiced with 177 Lu-prostate-specific membrane antigen and 177 Lu peptide receptor radionuclide therapy, cannot, therefore, be characterized as personalized precision medicine. The evolution of artificial intelligence (AI) could change this "one-size-fits-all" approach to theranostics, through development of a symbiotic relationship with physicians. Combining quantitative data collection, collation, and analytic computing power of AI algorithms with the clinical expertise, empathy, and personal care of patients by their physician envisions a new paradigm in theranostic reporting for molecular imaging and radioligand treatment of cancer. Human-AI interaction will facilitate the compilation of a comprehensive, integrated nuclear medicine report. This holistic report would incorporate radiomics to quantitatively analyze diagnostic digital imaging and prospectively calculate the radiation absorbed dose to tumor and critical normal organs. The therapy activity could then be accurately prescribed to deliver a preordained, effective, tumoricidal radiation absorbed dose to tumor, while minimizing toxicity in the particular patient. Post-therapy quantitative imaging would then valid…